The most commonly inactivated tumor suppressor gene in human cancer encodes the p53 protein, a transcription factor that is intimately involved in maintaining the integrity of the genome in a cell [Hall and Peters, Adv. Cancer Res., 68:67-108 (1996); Hainaut et al., Nucleic Acid Res., 25:151-157 (1997); Sherr, Cancer Res., 60:3689-95 (2000)]. In response to oncogenic stress signals, the cell triggers the p53 transcription factor to initiate either apoptosis or cell cycle arrest. Apoptosis facilitates the elimination of damaged cells from the organism, while cell cycle arrest enables damaged cells to repair genetic damage [reviewed in Ko et al., Genes & Devel. 10:1054-1072 (1996); Levine, Cell 88:323-331 (1997)]. The loss of the safeguard functions of p53 predisposes damaged cells to progress to a cancerous state. Inactivating p53 in mice consistently leads to an unusually high rate of tumors [Donehower et al., Nature, 356:215-221 (1992)].
The p53 transcription factor promotes the expression of a number of cell cycle regulatory genes, including the gene encoding the Mouse Double Minute (Mdm2) protein [see, Chene, Nature Reviews Cancer 3:102-109 (2003)]. The Mdm2 protein (designated Hdm2 in humans and Mdm2 in mice) acts to down-regulate p53 activity in an auto-regulatory manner [Wu et al, Genes Dev., 7:1126-1132 (1993); Bairak et al., EMBO J, 12:461-468 (1993)]. In the absence of oncogenic stress signals, i.e., under normal cellular conditions, the Mdm2 protein serves to maintain p53 activity at low levels [Wu et al, Genes Dev., 7:1126-1132 (1993); Barak et al., EMBO J, 12:461-468 (1993)].
Interestingly, whereas Mdm2 negative (Mdm2−/−) mice are not viable [Jones et al, Nature, 378:206-208 (1995); Montes de Oca Luna et al., Nature, 378:203-206 (1995)], additional inactivation of the p53 gene rescues Mdm2−/− mice [Jones et al, Nature, 378:206-208 (1995); Montes de Oca Luna et al., Nature, 378:203-206 (1995)]. These results indicate that the misregulation of the p53 transcription factor in the Mdm2 negative mice is the root cause of the observed lethality of the Mdm2−/− genotype, and that the regulation of p53 function relies on an appropriate balance between the two components of this p53-Mdm2 auto-regulatory system. Indeed, this balance appears to be essential for cell survival.
There are at least three ways that Mdm2 acts to downregulate p53 activity. First, Mdm2 can bind to the N-terminal transcriptional activation domain of p53 to block expression of p53-responsive genes [Kussie et al., Science, 274:948-953 (1996); Oliner et al., Nature, 362:857-860 (1993); Momand et al, Cell, 69:1237-1245 (1992)]. Second, Mdm2 shuttles p53 from the nucleus to the cytoplasm to facilitate the proteolytic degradation of p53 [Roth et al, EMBO J, 17:554-564 (1998); Freedman et al., Mol Cell Biol, 18:7288-7293 (1998); Tao and Levine, Proc. Natl. Acad. Sci. 96:3077-3080 (1999)]. Finally, Mdm2 possesses an intrinsic E3 ligase activity for conjugating ubiquitin to p53 within the ubiquitin-dependent 26S proteosome pathway [Honda et al., FEBS Lett, 420:25-27 (1997); Yasuda, Oncogene 19:1473-1476 (2000)]. Thus, Mdm2 impedes the ability of the p53 transcription factor to promote the expression of its target genes through binding p53 in the nucleus.
Attenuating the p53-Mdm2 auto-regulatory system can have a critical effect on cell homeostasis. Consistently, a correlation between the overexpression of Mdm2 and tumor formation has been reported [Chene, Nature 3:102-109 (2003)]. Since Mdm2 acts as a post-translational regulatory effector of the p53 transcription factor, compounds that hinder the ability of Hdm2/Mdm2, to interact with p53 would be anticipated to cause an immediate increase in p53 activity, and thereby rapidly promote either cell cycle arrest or apoptosis in damaged cells. Not surprisingly then, there is currently a substantial effort being made to identify new anticancer agents that hinder the ability of Hdm2 to interact with p53 [Chene, Nature 3:102-109 (2003)]. However, to date, no suitable anticancer agent has been found.
Structure-based drug design is one way to optimize the success of identifying useful antagonists of Hdm2, but use of this powerful methodology requires the three-dimensional structure of the target protein. So far, little information has been provided regarding the three-dimensional structure of Hdm2. Indeed, the only structures of Mdm2 currently available are those of Hdm2 and Xenopus Mdm2 (XMdm2), complexed with a p53 peptide, but neither crystalline form is suitable for structure-based drug design [Kussie, et al. Science, 274(5289): 948-953 (1996)]. Moreover, most of the protein-protein contacts in these crystal lattices are formed through the interaction of the exposed residues in the bound p53 peptide (D21, K24, and L25), making the p53 peptide difficult to displace, which makes it inaccessible to potential inhibitors.
In direct contrast, a successful structure-based drug design program focusing on Hdm2 requires a form of the Hdm2 protein that is amenable to crystallization in the absence of any particular binding partner. Further, the crystal form of the p53-binding pocket of the Hdm2 protein should be accessible to potential inhibitors used for testing binding or for co-structural determination. However, up until now, the solubility and stability of the free Hdm2 protein has been significantly less than that of the Hdm2-p53 peptide complex.
Thus, there is a need to obtain nucleic acids that encode an Hdm2 protein that is soluble and stable at high protein concentrations even when the protein is free of p53 or fragments thereof. In addition, there is a need to design purification procedures that lead to the preparation of an isolated active Hdm2 protein that is soluble and stable when independent of p53 or fragments thereof. In addition, there is a need to obtain reproducible crystals of Hdm2 that are of sufficient quality for X-ray crystallization analyses and structural determinations. There is also a need to provide methods for identifying inhibitors of Hdm2 through structure-based drug design and for combining potential inhibitors with the crystals of Hdm2 and analyzing their binding. Further, there is a need to obtain Hdm2 protein samples that, when combined with potential inhibitors, are amenable to forming homogenous crystals.